Author(s)

GIULIA MARTINO, VITTORIO ASTARITA, SINA SHAFFIEE HAGHSHENAS, GIUSEPPE GUIDO, SAMI SHAFFIEE HAGHSHENAS

Abstract

This paper presents an experiment conducted based on an emergency management system. The authors have designed the architecture of an emergency management system, integrating the Tritone microsimulation software with artificial intelligence technologies and real-time data collected through IoT devices. The goal is to develop a system capable of generating personalized evacuation routes for each user, based on the specific emergency in progress and their location within the territory. To assess the reliability and usefulness of the proposed system, an experiment was conducted simulating the collapse of a bridge in the heart of the industrial area of Rende (Italy), caused by a flood. This event rendered the shortest route to the assembly point indicated by the Italian Civil Protection unusable. The experiment examined users’ willingness to follow an alternative route suggested by a mobile application. The results showed that 90% of users would be willing to follow the instructions provided by the system via the app, after receiving appropriate training in its use. Based on this scenario, the present research used the Tritone software to simulate three different situations: in the first, all users follow the evacuation route suggested by the application; in the second, no user follows the guidance and everyone takes the route affected by the flood; in the third scenario, some of the users follow the route recommended by the app, while others choose the compromised path. The results of the three simulations are then compared, highlighting the benefits of an evacuation management system based on artificial intelligence and real-time updates.

Keywords

artificial intelligence, internet of things, digital twins, risk reduction, microsimulation software, disaster management